12
STRESS AND CONTINGENCY MANAGEMENT IN THE TREATMENT OF IRRITABLE BOWEL SYNDROME CONCEPCION FERNANDEZ, PhD 1 , MARINO PEREZ, PhD 1 , ISAAC AMIGO, PhD 1 AND ANTONIO LINARES, PhD 2 1 Dpto Psicologı´a, Universidad de Oviedo, Spain 2 Servicio de Gastroenterologı´a, Hospital Central de Asturias, Spain SUMMARY Ninety patients with irritable bowel syndrome (IBS) who did not respond to medical treatment were randomly assigned to four treatment conditions — two experimental groups: stress management and contingency management; and two control groups: medical treatment and placebo. The subjects underwent 12 individual sessions which were specific for each condition. All the subjects completed symptom-monitoring diaries. Thirty-three dropped out during the assessment or treatment. The subjects who received training in contingency management experienced significant reductions in all the characteristic digestive symptoms: abdominal pain p < 0:001, diarrhoea p < 0:05, constipation p < 0:05 and dyspepsia p < 0:001. At the end of the treatment, 50 per cent of the patients remained asymptomatic and 37.5 per cent reduced their symptoms by at least 50 per cent. Among the patients assigned to the condition stress management, 33 per cent got rid of their symptomatology and the subjects showed significant reductions in the following digestive symptoms: abdominal pain p < 0:05, diarrhoea p < 0:05 and dyspepsia p < 0:05. The changes in the placebo group are not representative. The subjects assigned to this condition showed a high dropout rate. Significant changes were not observed in symptomatology in the medical treatment group. The results are maintained after a year of follow-up. Possible predictive parameters of the progress of the patients are explored. # 1998 John Wiley & Sons, Ltd. Stress Med., 14: 31–42, 1998. KEY WORDS — irritable bowel syndrome; stress management; contingency management; behavioural model Irritable bowel syndrome (IBS) is a gastrointestinal disorder characterized by the continuous or inter- mittent presence of abdominal pain and alteration of bowel habits (diarrhoea and constipation) with the lack of a detectable organic cause. 1;2 Although it lacks mortality, the importance of this clinical chart stems from its high incidence. Between 10 and 20 per cent of the general population present symptoms compatible with IBS and between 50 and 70 per cent of the digestive patients receive an IBS diagnosis. 1;3–5 Medical treatment does not solve the problem for a high number of patients, who generally continue seeing a doctor. This presumes, besides personal discomfort, high costs in diagnosis proofs, in medicine consumption and worker absenteeism. 6;7 The implication of psychological factors is recognized in the onset and worsening of the symptomatology. 8 Usually patients report anxiety and depression although there is no relationship between IBS and specific psychological dis- orders. 9–11 Nor has it been possible to identify a specific personality profile of these patients except for higher values of neuroticism. 12–14 One charac- teristic, which is common to all IBS patients and which allows us to clearly dierentiate those with organic gastrointestinal alterations, is a behaviour pattern called ‘chronic illness behaviour’ 15 and which suggests the role that social learning could have in the aetiology of the disorder. 16–19 Latimer et al. 14 contrasted the physiological examinations of patients with IBS with those of other asymptomatic people but with the same Correspondence to: Dra Concepcio´n Ferna´ndez, Departa- mento de Psicologı´a, Universidad de Oviedo, Plaza Feijo´o s/ n. 33003 Oviedo, Spain. E-mail: [email protected]. Tel: 34-8-5103249. Fax: 34-8-5104144. Contract grant sponsor: DGICYT; Contract grant number: PS94/0146. CCC 0748–8386/98/010031–12$17.50 # 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31–42 (1998)

Stress and contingency management in the treatment of irritable bowel syndrome

Embed Size (px)

Citation preview

Page 1: Stress and contingency management in the treatment of irritable bowel syndrome

STRESS AND CONTINGENCY MANAGEMENTIN THE TREATMENT OF IRRITABLE

BOWEL SYNDROME

CONCEPCION FERNANDEZ, PhD1�, MARINO PEREZ, PhD1, ISAAC AMIGO, PhD1 AND ANTONIO LINARES, PhD2

1Dpto PsicologõÂa, Universidad de Oviedo, Spain2Servicio de GastroenterologõÂa, Hospital Central de Asturias, Spain

SUMMARY

Ninety patients with irritable bowel syndrome (IBS) who did not respond to medical treatment were randomlyassigned to four treatment conditions Ð two experimental groups: stress management and contingency management;and two control groups: medical treatment and placebo. The subjects underwent 12 individual sessions which werespeci®c for each condition. All the subjects completed symptom-monitoring diaries. Thirty-three dropped out duringthe assessment or treatment. The subjects who received training in contingency management experienced signi®cantreductions in all the characteristic digestive symptoms: abdominal pain � p < 0:001�, diarrhoea � p < 0:05�,constipation � p < 0:05� and dyspepsia � p < 0:001�. At the end of the treatment, 50 per cent of the patients remainedasymptomatic and 37.5 per cent reduced their symptoms by at least 50 per cent. Among the patients assigned to thecondition stress management, 33 per cent got rid of their symptomatology and the subjects showed signi®cantreductions in the following digestive symptoms: abdominal pain � p < 0:05�, diarrhoea � p < 0:05� and dyspepsia� p < 0:05�. The changes in the placebo group are not representative. The subjects assigned to this condition showed ahigh dropout rate. Signi®cant changes were not observed in symptomatology in the medical treatment group. Theresults are maintained after a year of follow-up. Possible predictive parameters of the progress of the patients areexplored. # 1998 John Wiley & Sons, Ltd.

Stress Med., 14: 31±42, 1998.

KEY WORDS Ð irritable bowel syndrome; stress management; contingency management; behavioural model

Irritable bowel syndrome (IBS) is a gastrointestinaldisorder characterized by the continuous or inter-mittent presence of abdominal pain and alterationof bowel habits (diarrhoea and constipation) withthe lack of a detectable organic cause.1;2 Althoughit lacks mortality, the importance of this clinicalchart stems from its high incidence. Between 10 and20 per cent of the general population presentsymptoms compatible with IBS and between 50and 70 per cent of the digestive patients receive anIBS diagnosis.1;3±5 Medical treatment does notsolve the problem for a high number of patients,who generally continue seeing a doctor. This

presumes, besides personal discomfort, high costsin diagnosis proofs, in medicine consumption andworker absenteeism.6;7

The implication of psychological factors isrecognized in the onset and worsening of thesymptomatology.8 Usually patients report anxietyand depression although there is no relationshipbetween IBS and speci®c psychological dis-orders.9±11 Nor has it been possible to identify aspeci®c personality pro®le of these patients exceptfor higher values of neuroticism.12±14 One charac-teristic, which is common to all IBS patients andwhich allows us to clearly di�erentiate those withorganic gastrointestinal alterations, is a behaviourpattern called `chronic illness behaviour'15 andwhich suggests the role that social learning couldhave in the aetiology of the disorder.16±19

Latimer et al.14 contrasted the physiologicalexaminations of patients with IBS with those ofother asymptomatic people but with the same

�Correspondence to: Dra Concepcio n Ferna ndez, Departa-mento de Psicologõ a, Universidad de Oviedo, Plaza Feijo o s/n. 33003 Oviedo, Spain. E-mail: [email protected]: 34-8-5103249. Fax: 34-8-5104144.

Contract grant sponsor: DGICYT; Contract grant number:PS94/0146.

CCC 0748±8386/98/010031±12$17.50# 1998 John Wiley & Sons, Ltd.

STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 2: Stress and contingency management in the treatment of irritable bowel syndrome

psychological pro®le Ð neuroticism. The authorsdid not ®nd di�erences either in electrical activityor in intestinal motility between both groups ofindividuals, the most signi®cant factor being thevariability among individuals. Latimer16 proposedan IBS behavioural model based on these datawhich would determine this entity as a learnedbehaviour inscribed inside a wider pattern ofneurotic behaviour. It is assumed, bearing in mindthe research related to the genetic predisposition toneuroticism,20 that problematic bowel responsescould be unlearned responses due to stress andcommon to other neurotic individuals, while verbalbehaviour and other symptomatic signs would belearned and would indicate the di�erences withother neurotic forms.

Several psychological interventions have beentried with IBS. Controlled studies have showndi�erent treatments based on tranquillizers andantidepressants to be equal to placebo conditionsin alleviating the symptoms of IBS. In this way,these bene®cial results have not been ascribed tospeci®c pharmacological e�ects.21±23 Hypnother-apy appears a more e�cient procedure, althoughits achievements could be attributed to the use ofrelaxation and to the doctor±patient relationshipitself.24 Finally, we have the behavioural inter-ventions: biofeedback, stress management andcontingency management.

The application of biofeedback procedures im-proves the control learned from bowel movementsalthough no clinical improvements are produced.25

The use of stress management procedures is justi-®ed, on the one hand, by the known worsening ofIBS under stressful situations and, on the otherhand, by the possibility of non-speci®c colonichyperreactivity as a response to events which are notobjectively stressful. The treatment modalities used,either combined or on their own, are relaxation,coping skills training and extinction proce-dures.26±30 Lastly, contingency management has todo with the operating character of illness manifesta-tions. It is a matter of decreasing social reinforce-ment of the symptomatic manifestations andsimultaneously focusing attention on conditionsof well-being.16 As a whole, improvement ratesshow the superiority of behavioural interventionsover medical treatments, in particular with thosepatients who do not respond to medical treatments.Otherwise, these procedures seem appropriate anduseful, at least, for a certain kind of patient.

Consequently, this research explores thesuperiority of behavioural therapy over other

procedures. To do this, those e�ects of thebehaviourist strategies that seem more viable, thatis, stress management and contingency manage-ment, are contrasted with the results of con-ventional medical treatment and with a placebocondition (in order to control the high responsive-ness of these patients to harmless therapies or otherinterfering variables such as the increase in atten-tion they are paid). Secondly, it attempts to identifythe most e�cient treatment and the clinical orbehavioural variables which would make onetreatment more appropriate than other.

METHOD

Subjects

A total of 90 patients with IBS were studied.Included in the research were subjects requestingmedical assistance in the Digestive System Serviceof CabuenÄ es Hospital (Asturias, Spain) who werediagnosed with IBS, had been su�ering from itfor over a year and presented, at least, two of thefollowing characteristics (indicative of a badprognosis according to a previous study31): wrongingestion of medicines prescribed by the doctorand/or uncompliance; more than one visit notscheduled by the gastroenterologist (after a pre-vious one in which the patient was told the natureof the illness); worker absenteeism or di�cultieswith carrying out ordinary job tasks as a con-sequence of the symptoms; previous psychiatrictreatment; and entry into the emergency servicewithout medical indication. The diagnosis wasmade based on the criteria of Manning et al.32 Inall cases, several complementary tests (haema-tology, biochemistry, including hepatic functiontests, upper gastrointestinal endoscopy and colono-scopy or contrast radiology, determination ofthyroid hormones, lactose tolerance test, faecalcultures, or barium examination of the smallintestine) were performed.

The medical clinic sample was drawn from100 patients. Ten of these patients refused toparticipate because of lack of time or no reasongiven, leaving 90 subjects who consented to theirparticipation in the study. Thirty-three droppedout during the treatment. The number of dropoutsper condition are listed in Table 1. The placebogroup reported more dropouts than the otherconditions.

The demographic data for the study sample�N � 90� are presented in Table 1. Comparisons of

32 C. FERNANDEZ ET AL.

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 3: Stress and contingency management in the treatment of irritable bowel syndrome

demographics across the four conditions revealedno signi®cant di�erences among conditions. More-over, treatment dropouts did not di�er fromcompleters on any variable.

Therapists

A total of four gastroenterologists took part inthe selection of the patients as blind examiners ofthe ®nal results.

Three psychologists with similar professionalexperience took part in the application of psycho-logical treatments. Each psychologist applied asingle treatment. The psychologist in charge of theplacebo condition did not know the nature of thistreatment in order to control for unfavourableexpectations of the experimenter interfering withthe results.

Assessment procedures

Behavioural interview. This was designed forthe purpose of gathering information related to

demographic and clinical questions, and to ®nd outabout the existence of symptomatology precipi-tants and reinforcement contingency. It was devel-oped in accordance with the A-B-C scheme thatde®nes the behavioural interview.

IBS symptom diary. At the conclusion of theinitial assessment visit, patients began keeping adaily digestive symptom diary. In it they recorded:(a) the presence and intensity of the symptoms(see list in Table 2), rating their severity on thefollowing subjective scale: not a problem � 0,mild � 1, moderate � 2, severe � 3, debilitat-ing � 4; (b) the subject and others' response todiscomfort; (c) other conditions such as: what he/she was doing, where, who with, how he/she felt.The aim was to identify the causality relationswhich explain the symptomatic manifestations.

Patients kept the IBS symptom diary for abaseline (2 weeks) and throughout treatment(nominally 10 weeks). The following commonsymptoms of IBS were rated: pain (de®ned interms of abdominal pain and discomfort relievedby evacuation or associated with change in the

Table 1 Ð Demographic characteristics of participants in each condition

Total Medico-C Placebo-C Stress-M Conting.-M p�N � 90� �N � 23� �N � 23� �N � 21� �N � 23�

Females (%) 66 69 66 66 75 NSMean age (yr) 44 46 49 47 40 NSEducation (%)

Elementary 68 72 80 77 69 NSMiddle school 17 24 10 23 26 NSHigh school 5 4 10 0 5 NS

Income (%)Low 39 38 37 44 42 NSMedium 58 62 60 53 60 NSHigh 3 0 3 3 8 NS

Profession (%)Housewife 43 48 47 47 50 NSUnquali®ed 36 40 37 30 30 NSTechnical 17 10 20 23 20 NSProfessional 4 2 6 0 0 NS

IBS history (yr) 8 6.7 6 5.9 7 NS

Average of visits per yearFamily doctor 3 3 3 3 3 NSGastroenterol. 3 3 3 3 3 NSOther doctors 2 2 1 2 2 NSEmergency 1 1 0 1 1 NS

Psychiatric treatment (%) 49 44 47 53 50 NSDropped out 33 4 16 6 7

STRESS AND CONTINGENCY MANAGEMENT IN IBS TREATMENT 33

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 4: Stress and contingency management in the treatment of irritable bowel syndrome

frequency or consistency of stools); diarrhoea(urgency, loose stools and increased frequency Ðmore than three daily); constipation (straining,incomplete evacuation and hard or infrequentstools Ð fewer than three weekly); dyspepsia(epigastric symptoms). The diary was completedtwice per day, at midday and in the evening. Forthe baseline and during the treatment (weekly), wecalculated average scores for each of the symptomsmonitored in the IBS symptom diary for eachpatient.

Treatment procedures

Once the subjects had been selected by thegastroenterologists they were randomly distributedto each of the treatment conditions described below(Table 1). The application of the treatments con-sisted of 12 weekly individual sessions of approxi-mately 1 h each, led by the same therapist. The ®rsttwo sessions comprised the psychological evalua-tion of each patient, the establishment of thebaseline of the symptomatology and brief educa-tion about normal bowel function, in line withLatimer's.16 The 10 sessions left were speci®c toeach condition.

1. Control group: conventional medical treat-ment. Patients in this group did not under-go a weekly visit programme but receivedconventional medical treatment. They werereassured of the e�ectiveness of standard drugtherapy.

2. Placebo control group. In designing the place-bo group, we sought a treatment which would:be credible, generate a high level of expecta-tion and bene®t and be psychologicallyengaging while at the same time having nopsychological or clinical e�ects. The proceduresconsisted of some imaginative and activevisualization of bowel function exercises andthe prompting of their own capacity for self-regulation through thought, stimulating theirconcentration to the utmost. They were nottrained in relaxation. The patient was asked tovisualize in his body `in red' the area a�ectedby his/her symptomatology. Afterwards, hewas given a glass of water and was suggestedthe image of `a blue liquid that goes down thedigestive system cleansing the red colourand eliminating the discomforting sensations'.Patients were asked to practise the exercise ontheir own at home at least once per day and to

complete their self-reports. The 10 experi-mental sessions were used in the practiceof this exercise. The therapist maintainsenthusiasm all the time and encourages thesubject to practise daily.

3. Experimental group: stress management. Theaim of this condition was to provide thepatient with e�ective techniques to mitigatethe physiological e�ects of stress and ten-sion, and to modify his/her coping skills.To achieve this, patients were trained inprogressive muscular relaxation according tothe programme described by Bernstein andBorkovec;33 this began with 16 muscle groups,which reduced to eight, and then four. Therelaxation training was taught over the ®rstsix sessions. Relaxation by recall and cue-controlled relaxation were also taught. Anaudiotape was given to the patients to guideregular home practice, which was encouraged.Patients were asked to practise daily for about20 min and a report of achievements wasrequired. Training in stress management wasintroduced in session 7 and was carriedout according to the procedure described byMeichenbaum,34 using relaxation practice, self-instruction, problem-solving and coping ineveryday life with experiences related to theonset or worsening of the symptomatology.

4. Experimental group: contingency management.The aim of this therapy modality was to showthe patient and, optionally and whenever it waspossible, other people relevant to his/her every-day life (family, friends) how to practisebehaviours more adapted to IBS symptomsand which situations made them worse and,alternatively, to extinguish inadequate beha-viours (i.e. isolation, dependence, delegation,overprotection) of the patient or of others in thepresence of those symptoms. The basic instru-ment for the application of the treatment wasthe contingency contract devised according tothe speci®cations of O'Banion and Whaley.35

The new behaviours to be established, speci®edin operative terms, and the reinforcementcontingencies to be undergone are agreedbetween the patient and his/her relatives orother people relevant to the patient and thetherapist. Other behavioural techniques usedduring the treatment were self-observation andshaping, stimulus control to neutralize inade-quate habits or to break learned illness beha-viour chains, and the restructuring of time and

34 C. FERNANDEZ ET AL.

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 5: Stress and contingency management in the treatment of irritable bowel syndrome

training in social skills in order to increase therate of alternative rewards to the symptomaticmanifestations.

Post-treatment assessment and follow-up

Once the intervention was ®nished, a gastro-enterologist, who did not know the status of thesubjects, evaluated the clinical change of eachpatient on the following scale: asymptomatic(no symptoms); improvement (all symptom averagescores were 550 per cent with respect to BL);unimprovement (no change); worsening (two ormore symptom average scores were 450 per centwith respect to BL).

For the analysis of data, the scores for eachdigestive symptom monitored in the symptomdiary obtained during a session just after thetreatment were considered as post-treatmentscores.

The follow-up of the patients after a year wasdone by means of a telephone call in which theywere asked if they had any digestive discomfort, ifthey were following the medical treatment, if theyhad been back to the gastroenterologist since thelast check-up and if they had su�ered any episodeof intense pain.

DATA ANALYSIS

Digestive symptoms monitored in the symptom diary

Analysis of covariance, with pre-treatmentscores of digestive symptoms as the covariate, wascarried out. Student's t-test was used for the posthoc comparisons.

Clinical change

A comparative analysis (using chi-square)among groups of the proportion of patients whereeach condition had achieved a clinically signi®cantchange was made.

Illness behaviour

The presence before and after treatment of thefollowing conditions was evaluated: psychiatrictreatment, adhesion to the medical treatment,anxiety, symptomatology precipitants, social atten-tion related to the symptoms, verbal and motormanifestations of illness (complaints, pain gestures,quitting activities, work delegation, increase ofmanifestations in front of relatives or friends,

avoiding situations associated with the presenceof symptoms). With each variable we proceeded tocompare (using chi-square) the pre±post treatmentincidence between and intra groups. Moreover, acomparison of the percentages (using chi-square)among the subjects who experienced recovery inthe post-treatment with those who did not pro®tfrom the intervention and among the patients who®nished the treatment and those who dropped outwas made for each of the variables.

Behavioural predictors

The correlation coe�cients among the variablesdemographic, treatment conditions, illness beha-viours and clinical improvement were calculatedand a discriminant analysis among all the variablesusing the clinically signi®cant improvement ascriterion was made in order to obtain the possiblepredictive parameters of the patients' progress.

RESULTS

Digestive symptoms

Table 2 shows the average pre-treatment andpost-treatment scores for each of the IBS symp-toms monitored in the symptom diary.

Analysis of covariance showed that the groupswere similar with respect to abdominal pain,constipation, diarrhoea and dyspepsia as measuredat referral during the ®rst baseline session.

Abdominal pain. Analysis of covariance revealeda treatment group di�erence on abdominal pain�F � 4:12; p < 0:032� following treatment. Sub-sequent analysis revealed signi®cant di�erencesbetween the contingency management group andthe medical control group � p < 0:001�; between thestress management group and the medical controlgroup � p < 0:022�; whereas no di�erences werefound between the contingency management groupand the stress management group, or between thecontingency management group and the placebogroup or the stress management group and theplacebo group. A Student's t-test showed thatabdominal pain decreased signi®cantly in the con-tingency management group � p < 0:001�, the stressmanagement group � p < 0:05� and the placebogroup � p < 0:01�.

Diarrhoea. Analysis of covariance revealed atreatment group di�erence on diarrhoea �F � 3:12;

STRESS AND CONTINGENCY MANAGEMENT IN IBS TREATMENT 35

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 6: Stress and contingency management in the treatment of irritable bowel syndrome

p < 0:048� following treatment. Subsequentanalysis revealed signi®cant di�erences betweenthe contingency management group and themedical control group � p < 0:001�; between thestress management group and the medical controlgroup � p < 0:001�; between the contingencymanagement group and the placebo group� p < 0:05�; and between the stress managementgroup and the placebo group � p < 0:05�; whereasno di�erences were found between the contingencymanagement group and the stress managementgroup, or between the placebo group and themedical control group. A Student's t-test showedthat diarrhoea decreased signi®cantly in the con-tingency management group � p < 0:05� and thestress management group � p < 0:05�.

Constipation. Analysis of covariance revealeda treatment group di�erence on constipation�F � 2:3; p < 0:10� following treatment. Subse-quent analysis revealed signi®cant di�erencesbetween the contingency management group andthe medical control group � p < 0:05�; between thecontingency management group and the stressmanagement group � p < 0:05�; and between thecontingency management group and the placebogroup � p < 0:05�; whereas no di�erences werefound between the stress management group andthe medical control group, or between the stressmanagement group and the placebo group or theplacebo group and the medical control group. A

Student's t-test showed that constipation onlydecreased signi®cantly in the contingency manage-ment group � p < 0:01�.

Dyspepsia. Analysis of covariance revealed atreatment group di�erence on dyspepsia �F � 5:12;p < 0:02� following treatment. Subsequent analysisrevealed signi®cant di�erences between the con-tingency management group and the medicalcontrol group � p < 0:001�; between the contin-gency management group and the placebo group� p < 0:05�; between the stress management groupand the medical control group � p < 0:05�; andbetween the stress management group and theplacebo group � p < 0:05�; whereas no di�erenceswere found between the contingency managementgroup and the stress management group, or theplacebo group and the medical control group. AStudent's t-test showed that dyspepsia decreasedsigni®cantly in the contingency management group� p < 0:001� and in the stress management group� p < 0:05�.

Clinically signi®cant improvement

Table 3 shows the clinical change for each ofthe conditions. As regards the improvement rateof patients who ®nished the intervention, thecontingency management group reached the high-est rate of asymptomatic subjects (50 per cent), aswell as signi®cant reduction of symptomatology

Table 2 Ð Pre-treatment and post-treatment digestive symptom scores

Symptom Condition Pre-treat. SD Post-treat. SD pX X

Abdominal pain Medico-C 16.1 8.3 15.8 9.9 NSF � 4:12, Placebo-C 15.3 9.3 2.3 9.8 0.01p � 0:032 Stress-M 15.7 10.9 6.8 5.5 0.05

Conting.-M 14.6 7.1 3.1 5.8 0.001

Diarrhoea M-C 13.7 8.6 15.2 5.2 NSF � 3:12, P-C 9.2 7.8 9.9 5.9 NSp � 0:048 S-M 13.2 8.3 6.3 9.7 0.05

C-M 12.8 9.5 5.2 6.1 0.05

Constipation M-C 12.5 12.4 11.3 10.3 NSF � 2:33, P-C 9.4 10.4 10.1 8.8 NSp � 0:10 S-M 9.8 9.2 9.5 5.9 NS

C-M 11.1 12.2 6.2 7.9 0.05

Dyspepsia M-C 15.4 12.0 14.5 13.2 NSF � 5:12, P-C 10.7 10.2 10.2 8.9 NSp � 0:028 S-M 13.9 9.4 5.8 6.4 0.05

C-M 14.1 13.4 4.6 6.9 0.001

36 C. FERNANDEZ ET AL.

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 7: Stress and contingency management in the treatment of irritable bowel syndrome

(37.5 per cent). It is observed in the stress manage-ment group that after the intervention 33.3 per centof the subjects remained asymptomatic and20 per cent had decreased symptomatology. Therewas an improvement in ®ve out of seven patientsfrom the placebo group who ®nished the pro-gramme. On the other hand, changes related to thepre-treatment were not observed in the medicalcontrol group. We compared the scores of the fourgroups and found signi®cant di�erences among thefour groups (chi-square� 26:17�9�, p < 0:01�.

Follow-up

Table 4 shows the patients' clinical conditions inthe follow-up in relation to the post-treatmentclinical situation (information was not available for15 subjects). The substantial symptomatic reliefobtained at post-treatment appeared to be main-tained at 12-month follow-up. After a year, 42.01per cent of the patients had experienced no changeover post-treatment digestive symptoms levels;14.03 per cent rated themselves as improved by atleast 50 per cent when compared with their IBS

status post-treatment; while 17.55 per cent wereworse with respect to their post-treatment status.These data reveal good maintenance of symptom-atic improvements over a 12-month follow-up(chi-square� 13:18�4�; p < 0:05�.

Illness behaviour

Table 5 shows the incidence of illness behaviourin the pre-treatment of those subjects whoimproved post-treatment, unimproved post-treatment and dropped out of the treatment, byeach condition.

Psychiatric treatment. 44.4 per cent of the totalsample received, previous to the study, psychiatrictreatment without di�erences among the groups. Inthis variable there are no pre-treatment signi®cantdi�erences among those patients who ®nished thetreatment and those who dropped out. Thosetreated psychiatrically before that intervention aresigni®cantly more numerous � p < 0:05� amongthe non-recovered. These data, coinciding withthe ones referred to in the bibliography, seem to

Table 3 Ð Clinical change post-treatment

Condition Asymptomatic Improved Unimproved Worse

Medical-control �N � 19� 1 2 10 6Placebo-control �N � 7� 1 4 1 1Stress-manag. �N � 15� 5 3 6 1Contingency-manag. �N � 16� 8 6 1 1

Total 26.37% 26.37% 31.57% 15.78%

Table 4 Ð Patients' clinical conditions in follow-up in relation to post-treatment

No change follow-up Improved follow-up Worse follow-up Unknown(42.1%) (14.03%) (17.5%) (26.33%)

No change post-treatmentM-C (10) 2 3 1 4P-C (1) 1 Ð Ð ÐS-M (6) 2 3 1 ÐC-M (1) 1 Ð Ð Ð

Improved post-treatmentM-C (3) 1 Ð 2 ÐP-C (5) 1 Ð 3 1S-M (8) 5 Ð 1 2C-M (14) 7 Ð 2 5

Worse post-treatmentM-C (6) 2 1 Ð 3P-C (1) 1 Ð Ð ÐS-M (1) Ð 1 Ð ÐC-M (1) 1 Ð Ð Ð

STRESS AND CONTINGENCY MANAGEMENT IN IBS TREATMENT 37

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 8: Stress and contingency management in the treatment of irritable bowel syndrome

support the predictive value of the bad prognosisassociated with the psychiatric treatment.

Adherence to the medical treatment of IBS. Thepatients were questioned regarding the completionof the treatments prescribed before the study;50 per cent of them revealed not having taken themedicines correctly (whether concerning dosage orin the recommended time). This variable, that hasbeen proposed by Linares-RodrõÂ guez as a para-meter of a bad prognosis, in our study is associatedneither with participation in the treatment nor withthe patient's progress. However, this conclusion isnot established as de®nite since its investigationthrough the patient's report is subject to importantconfounding factors such as social desirability.

Anxiety. A functional analysis of the responsesof anxiety evaluated in the interview and the self-reports was made. In order to determine thepresence of anxiety behaviours it was requiredfor patients to manifest anxiety behaviours daily(or several times a week) verbally, motorwise orphysiologically, and that these would interfere withand/or mean a subjective loss of control in theireveryday activities. However, it was not consideredwhether such responses were directly related to the

occurrence of symptoms. Among all the subjects inthe sample, 64 per cent of the patients showedanxiety during the baseline. Before the interventionthere were no di�erences in the incidence of thisvariable among the treatment groups. There wereno pre-treatment di�erences among the treatedpatients and those who dropped out, nor amongthe recovered or non-recovered patients on thisvariable. In the post-treatment, only those whoreceived training in stress management � p < 0:05�and also contingency management � p < 0:001�signi®cantly reduced anxiety responses.

Symptomatology precipitants. The subject's owncondition and/or the environment which caused orwas directly associated with the occurrence of thesymptomatology were identi®ed by means of aninterview or self-reports. The presence of precipi-tating stimulants during the baseline a�ected43 per cent of all the subjects. At the pre-treatmentthere were no di�erences among the groups.Among those precipitants identi®ed as physio-logical, the most often referred to were thefollowing: insomnia (18 per cent of the sample),excitement or nervousness (15 per cent), somefoods (12 per cent), bowel noises or movements(10 per cent), tiredness (8 per cent). Among those

Table 5 Ð Illness behaviour in the pre-treatment and patients' clinical conditions in the post-treatment

Psychiatric Adherence Anxiety Symptom Social Illnesstreatment medical treat. precipitant attention behaviourY N Y N Y N Y N Y N Y N

Improvement post-treatmentM-C 0 1 1 0 0 1 0 1 1 0 0 1P-C 2 0 1 1 2 0 1 1 2 0 0 2S-M 1 4 2 3 6 1 4 1 0 5 0 5C-M 3 7 6 4 9 1 8 2 7 3 8 2

Unimprovement post-treatmentM-C 11 7 10 8 11 7 8 10 10 8 9 9P-C 0 1 1 0 0 0 1 0 1 0 0 1S-M 5 2 1 6 6 1 5 2 4 3 5 2C-M 4 0 2 2 3 1 2 2 3 1 4 0

Dropped outM-C 1 3 2 2 1 3 1 3 2 2 2 2P-C 6 6 10 2 11 1 2 10 9 3 8 4S-M 3 2 2 3 3 2 2 3 2 3 2 3C-M 2 4 3 3 4 2 2 4 3 3 5 1

Improved/unimpovedChi2 5:1; p < 0:05 0.05 NS 0.65 NS 1.30 NS 0.12 NS 9:2; p < 0:01

Finished treatment/dropped outChi2 0.03 NS 1.30 NS 0.33 NS 7:6; p < 0:01 0.30 NS

38 C. FERNANDEZ ET AL.

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 9: Stress and contingency management in the treatment of irritable bowel syndrome

identi®ed as environmental were: family quarrels(21 per cent), low activity (14 per cent), workproblems (14 per cent), excessived work of familydemands (7 per cent), routine changes (7 per cent),cancerophobia (7 per cent), decision-taking(5 per cent), unknown situations (4 per cent) andworking shifts (3 per cent). Nevertheless, it shouldbe made clear that there were few patients in whomit was possible to identify safe and well-de®nedsymptom-precipitating events. In general, thesymptoms occurred in less speci®c situations inwhich the social e�ciency of the patient wasinvolved. Among those who ®nished the treatmentthe proportion of subjects in whom precipitants inthe pre-treatment were observed was signi®cantlyhigher than the dropouts. There were no pre-treatment di�erences among the recovered or non-recovered patients. Among treatment conditions,only stress management � p < 0:05� and contin-gency management � p < 0:05� were useful inreducing the presence of precipitants among theassigned subjects.

Social attention. It was decided that a patientwould receive social attention in the presence of asymptom when he got help, care or simply verbalattention and/or when verbal or motor manifesta-tions of pain occurred with a higher probabilityor were increased in the presence of relatives orfriends. Forty-one per cent of the sample receivedsocial attention during the baseline without di�er-ences among the groups. There were no pre-treatment di�erences among the treated patientsand those who dropped out, nor among therecovered or non-recovered patients on this vari-able. The attention that the subjects got after thetreatment was only modi®ed in the contingencymanagement group. However, among non-recov-ered patients in this same treatment condition therewere no changes, even when this was one of theobjectives of the intervention. In the stress manage-ment group, the same situation as the one beforethe treatment was maintained.

Learned illness behaviour. The presence oflearned illness behaviour characterized the studysample and there were no di�erences among thegroups. All the patients maintained some illnessbehaviour and 60 per cent of the subjects displayedthree or more of the six manifestations evaluated(see data analysis). The presence of learned illnessbehaviour before treatment did not allow us todi�erentiate between the patients who ®nished and

those who dropped out of the treatment. It did,however, allow us to do so among recovered andnon-recovered ones. In the contingency manage-ment group, they were signi®cantly reduced amongthose who progressed favourably � p < 0:01�. How-ever, in the stress management group, none of therecovered subjects presented a consolidated chronicillness behaviour before the treatment. Only thetraining in contingency management was able toe�ectively modify this behaviour, and in thosecases where this did not occur, reductions in thesymptomatology of IBS were not achieved.

Behavioural predictors

The data so far discussed suggest both thesuitability of the behavioural intervention andthe existence of certain parameters associated withthe clinical progress of the patients. In fact, thediscriminant analysis obtained from the datareferred to so far out of a total of 89 patients(one was excluded due to lack of data) permits us tocorrectly classify 82.02 per cent of the sample inrelation to their clinical improvement based onthe following variables: sex, age, psychiatric treat-ment, IBS duration, precipitant symptoms, socialattention, illness behaviours, family cooperation intreatment, stress and contingency managementtraining. The variables with the highest weightavailable at the time to predict the patient'sdigestive progress were `family's participation intreatment' (0.51) and `contingency managementtraining' (0.48). `Anxiety' and the presence of the`precipitant symptomatology' were positivelyrelated to the discriminatory function for digestiveimprovements. However, `illness behaviours' and`social attention' had a negative correlation.

DISCUSSION

The hypothesis and design of this research supportthe learned character of the IBS. The behaviouralpattern of the IBS can explain satisfactorily theacquisition and maintenance of the physiological,motor and cognitive manifestations as conditionedresponses by means of a paradigm of eitherPavlovian conditioning type 2 or operant con-ditioning.16 The validity of the behavioural patternrequires testing the e�ectiveness of the derivedpsychological interventions. Thus, anxiety andstress control techniques would seem appropriateand useful, at least, for those patients whose

STRESS AND CONTINGENCY MANAGEMENT IN IBS TREATMENT 39

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 10: Stress and contingency management in the treatment of irritable bowel syndrome

symptomatology is aggravated under conditionsof extra burden or emotional change, whereascontingency management would be justi®ed inthose cases where symptomatic occurrence appearsdepending on certain circumstances and is main-tained by attention or privileges.

In light of these considerations, this study teststhe di�erent e�cacy of the stress management andcontingency management techniques with regardto the conventional medical treatment of the IBSand a placebo condition. Moreover, it attempts to®nd out which aspects of the treatments or of thesubjects would make a psychological interventionespecially appropriate.

The results relating to the demographic data andthe clinical case histories of the patients allow us toassume that as regards these variables the studiedsample was representative of the IBS-a�ectedpopulation as it is described in the literature.

The results concerning the evolution of theclinical symptomatology are very telling. Whilebefore the treatment the subjects did not di�ereither statistically or in the intensity or frequencyof the discomfort, after the treatment we onlyobtained very signi®cant di�erences � p < 0:001� inthe contingency management group. Consideringthe symptoms individually, all the characteristicdigestive symptoms were signi®cantly � p < 0:05�reduced with training in contingency manage-ment training. The subjects' own estimation ofthe development of their illness indicated that50 per cent of them remained asymptomatic afterthe treatment and 37.5 per cent reduced theirdiscomfort.

The e�cacy of contingency management in thetreatment of IBS is obvious, and particularly withregard to the conventional medical treatment,where statistically signi®cant reductions were notobtained. At the end of the treatment, only15.7 per cent of the subjects considered that thediscomfort had been reduced, while 52.6 per centreported unimprovement and 31.5 per cent hadincreased discomfort.

In the placebo control groups, the results showedthat ®ve out of seven patients who ®nished thetreatment sessions improved. It should be notedthat 69 per cent of the subjects assigned to thiscondition dropped out. On the other hand,knowing the response of these patients to harmlesstreatments, these improvements are easily under-standable. Therefore, in order to evaluate theusefulness of this contrast group, it would be moreinteresting to explain the high dropout.

The low adherence to the treatments directlya�ects their generalization. Thus, it is important toinquire into the variables related to dropping out.The subjects who dropped out of this placebocondition did so on average in the fourth session,while those who abandoned other treatmentconditions did so on average in the ®rst session.The fourth was also the average session in whichthe beginning of improvement in patients whoprogressed favourably was signi®cantly present.The improvement can be attributed to the treat-ments because patients who ended the two experi-mental and the placebo treatments reported in theirdiaries an average of at least ®ve home practices perweek throughout treatment. The two experimentalgroups did not di�er in average home practice perweek. Due to dropout in the placebo group,comparisons with the other conditions are notadequate; in any case, the average in this group wasalso ®ve home practices per week. Thus it could beassumed that this dropout in placebo conditionswas related to the poor e�ectiveness of theprocedure in alleviating the discomfort. The lowcredibility of this treatment could itself also be afactor explaining the dropout. In order to approachthis problem, other studies have used question-naires to assess the credibility of the treatment andthe expectations of the patients before and afterit. In any case, adherence to the treatment couldnot be understood without the individual learninghistory of the illness and its treatment of eachsubject. In this study, the variables that bestpredicted dropout were: age over the groupaverage, wide repertoire of learned illness beha-viours, social attention in the presence of symp-toms, lack of anxiety and no collaboration of thefamily in the treatment. As a whole, it seems thatsubjects who display functional illness behavioursin their social and family context have a greaterpossibility of quitting or progressing unfavourably.In the stress management group, after the treat-

ment patients only obtained very signi®cant di�er-ences � p < 0:05� in abdominal pain, diarrhoeaand dyspepsia, although they did not reportchanges in constipation either. Nevertheless, inthe post-treatment 40 per cent of subjects reportedunimprovement, 33 per cent appeared asympto-matic and 20 per cent had reduced their symptoms.It is assumed that this important improvement in agroup of subjects could be explained by thepatients' individual characteristics.

Thus, and agreeing with the initial approach,it would appear likely that training in stress

40 C. FERNANDEZ ET AL.

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 11: Stress and contingency management in the treatment of irritable bowel syndrome

management would be especially useful to controlIBS manifestations in those patients with anxietyand/or those in whom symptom precipitantswere observed and who displayed an avoidancebehaviour which decreased or made the symptomsdisappear. In the pre-treatment interview it wasobserved that 64 per cent of the sample showedphysiological, motor and cognitive anxietyresponses. As regards the presence of stimulantsthat precipitate symptoms, they could also beidenti®ed during the baseline in 43 per cent of thesubjects. Anxiety responses and the existence ofspeci®c conditions that were symptom precipitantswere found in the pre-treatment in all the sub-jects who improved when receiving training instress management. However, in no case didthese patients show a consolidated learned illnessbehaviour nor did they receive social and/ormaterial attention in the presence of their sympto-matic manifestations.

The presence of learned illness behaviours hasappeared as a very signi®cant parameter in theprogress of patients. Thus, while in the stressmanagement group other patients who also showedanxiety and symptom precipitants in the baselineafter the treatment were able to solve this problem(as might be predicted), they did not produceimprovements in symptomatology. In all thesecases, patients (before the intervention) alsoshowed a consolidated chronic illness behaviour.In the contingency management group, all the non-recovered subjects continued obtaining gains intheir alterations after the treatment (even when thisitself was the objective of the intervention). How-ever, they were not maintained among those whorecovered. Contingency management was alsoe�ective in reducing the anxiety of patients inthose cases where the responses were concernedwith non-speci®c situations, generally of a socialcharacter, where the skills of the subject themselveswere involved (decision-taking, work overload,etc). It is understood that contingency manage-ment is e�ective in solving this problem preciselybecause the subject is provided with resources togive a functional and alternative response to thesymptom in the situations in which it is produced.Finally, the coincidence of anxiety and symptom

precipitants is associated with a favourable prog-nosis when the patient's symptoms are not paidattention to. In the cases in which those learnedillness behaviours do appear, the predictive valueof this parameter changes depending on othervariables with which it is associated. Thus when

learned illness behaviours are present andthe family takes part in the treatment and there istraining in contingency management, a favour-able outcome is predicted. The cases with theworst prognosis comprise psychiatric treatment,advanced age, learned illness behaviour andobtaining social attention.

In conclusion, the obtained data allow theconclusion that IBS consists mainly of the acqui-sition of a maladjusted behaviour and thattechniques of behaviour modi®cation allow thedisappearance or diminution of the patient'schronic illness behaviour and, therefore, thereduction of the number of visits, work absentee-ism and the patient's personal and social disorder.Further investigation is required to con®rm therelevance of the described predictors and to maketheir use pro®table for the gastroenterologisthimself.

ACKNOWLEDGEMENT

This research was partially ®nanced by a grantfrom DGICYT (PS94/0146).

REFERENCES

1. Thompson, W. G. The irritable bowel syndrome.Gut 1984; 25: 305±320.

2. Drossman, D. A., Thompson, W. G., Talley, N. J.,Funch-Jensen, P. and Whitehead, W. E. Identi®ca-tion of subgroups of functional gastrointestinaldisorders. Gastroenterology 1990; 3: 159±172.

3. Thompson, W. G. and Heaton, H. W. Functionalbowel disorders in apparently healthy people.Gastroenterology 1980; 79: 283±288.

4. Drossman, D. A., Sandler, R. S., McKee, D. C.and Lovitz, A. J. Bowel patterns among subjectsnot seeking health care. Use of a questionnaireto identify a population with bowel dysfunction.Gastroenterology 1982; 83: 529±534.

5. Mitchell, C. M. and Drossman, I. A. Survey ofthe AGA membership relating to patients withfunctional gastrointestinal disorders. Gastroenter-ology 1987; 92: 1282±1284.

6. Kingham, J. and Dawson, A. M. Origin of chronicright upper quadrant pain. Gut 1985; 26: 783±788.

7. Harvey, R. F., Manuad, E. C. and Broun, A. M.Prognosis in the irritable bowel syndrome: A 5-yearprospective study. Lancet 1987; 25: 963±965.

8. Switz, D. M. What the gastroenterologist does allday? Gastroenterology 1976; 70: 1048±1050.

9. Macdonald, A. J. and Bouchier, I. A. Non-organicgastrointestinal illness: A medical and psychiatricstudy. Brit. J. Psychiat. 1980; 136: 276±283.

STRESS AND CONTINGENCY MANAGEMENT IN IBS TREATMENT 41

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)

Page 12: Stress and contingency management in the treatment of irritable bowel syndrome

10. Craig, T. K. and Brown, G. W. Goal frustration andlife events in the aetiology of painful gastrointestinaldisorders. J. Psychosom. Res. 1984; 28: 411±421.

11. Welch, G. W., Hillman, L. C. and Pomare, E. W.Psychoneurotic symptomatology in the irritablebowel syndrome: A study of reporters and non-reporters. Brit. Med. J. 1985; 291: 1382±1384.

12. Palmer, R. L., Stonehill, E., Crips, A. H., Walker,S. L. and Misiewicz, J. J. Psychological character-istics of patients with the irritable bowel syndrome.Postgrad. Med. J. 1974; 50: 416±419.

13. Esler, M. D. and Goulston, H. T. Levels of anxiety incolonic disorders.N. Engl. J. Med. 1973; 288: 16±20.

14. Latimer, P. R., Sarna, S. K. and Campbell, D.Colonic motor and mioelectrical activity: A com-parative study of normal subjects, psychoneuroticpatients and patients with irritable bowel syndrome.Gastroenterology 1981; 80: 893±901.

15. Whitehead, W. E., Winget, C., Fedoravicius, A. S.,Wooley, S. and Blackwell, B. Learned illnessbehaviour in patients with irritable bowel syndromeand peptic ulcer. Dig. Dis. Sci. 1982; 27: 202±207.

16. Latimer, P. R. Functional Gastrointestinal Disorders.A Behavioural Medicine Approach. Springer, NewYork, 1983.

17. Sandler, R. S., Drossman, D. A., Nathan, H. P. andMcKee, D. C. Symptom complaints and health careseeking behaviour in subjects with bowel dysfunc-tion. Gastroenterology, 1984; 87: 314±318.

18. Drossman, D. A., Mckee, D. C., Sandler, R. S.,Mitchell, M., Cramer, E., Lowman, B. and Burger,A. Psychosocial factors in the irritable bowelsyndrome. Gastroenterology 1988; 95: 701±708.

19. Smith, R. C., Greenbaum, D. S., Vancouver, J. B.,Henry, R. C., Reinhart, M. A., Greenbaum, R. B.,Dean, H. A. and Mayle, J. E. Psychosocial factorsare associated with health care seeking rather thandiagnosis in irritable bowel syndrome. Gastroenter-ology 1090; 98: 293±301.

20. Eysenck, H. J. and Broadhurst, P. L. Experimentswith animals. In: Experiments in Motivation.Eysenck, H. I. (Ed.) Pergamon, Oxford, 1964.

21. Steinhart, M., Wong, P. I. and Zarr, M. L.Therapeutic usefulness of amitriptyline in spasticcolon syndrome. Int. J. Psychiatr. Med. 1981; 11:54±57.

22. Shapiro, A. K., Struening, E. L., Shapiro, E. andMilcarek, B. Diazepam: How much better thanplacebo? J. Psychiatr. Res. 1983; 17(1): 51±73.

23. Creed, F. andGuthrie, E. Psychological factors in theirritable bowel syndrome. Gut 1987; 28: 1307±1318.

24. Whorwell, P. J., Prior, A. and Colgan, S. M.Hypnotherapy in severe irritable bowel syndrome:Further experience. Gut 1987; 28: 423±425.

25. Whitehead, W. E. Psychotherapy and biofeedbackin the treatment of irritable bowel syndrome.In: Irritable Bowel Syndrome. Read, N. W. (Ed.)Grune and Stratton, London, 1985.

26. Ne�, D. F. and Blanchard, E. B. A multi-componenttreatment for irritable bowel syndrome. Behav. Ther.1987; 18: 70±83.

27. Blanchard, E. B., Schwarz, S. P. and Ne�, D. F.Two-year follow-up of behavioural treatmentof irritable bowel syndrome. Behav. Ther. 1988; 19:67±73.

28. Lynch, P. N. and Zamble, E. A controlled behavioraltreatment study of irritable bowel syndrome. Behav.Ther. 1989; 20: 509±523.

29. Schwarz, S. P., Taylor, A. E., Schar�, L. andBlanchard, E. B. A four-year follow-up of beha-viourally treated irritable bowel syndrome patients.Behav. Res. Ther. 1991; 28: 331±338.

30. Blanchard, E. B., Schwarz, S. P., Suls, J. M.,Gerardi, M. A., Schar�, L., Greene, B., Taylor,A. E., Berreman, Ch. and Malamood, S. Twocontrolled evaluations of multicomponent psycho-logical treatment of irritable bowel syndrome. Behav.Res. Ther. 1992; 30: 175±189.

31. Linares-RodrõÂ guez, A. Irritable bowel syndrome:epidemiology and treatment. Doctoral Thesis,University of Oviedo, Spain, 1986.

32. Manning, A. P., Thompson, W. G., Heaton, K. W.and Morris, A. F. Towards positive diagnosis ofthe irritable bowel syndrome. Brit. Med. J. 1978; 2:653±654.

33. Bernstein, D. A. and Borkovec, T. ProgressiveRelaxation Training. Research Press, New York,1973.

34. Meichenbaum, D. Stress Inoculation Training.Pergamon, New York, 1985.

35. O'Banion, D. R. and Whaley, D. L. Behaviour Con-tracting. Springer, New York, 1981.

42 C. FERNANDEZ ET AL.

# 1998 John Wiley & Sons, Ltd. STRESS MEDICINE, VOL. 14: 31±42 (1998)